--- Log opened Mon Jan 30 00:00:29 2023
00:11 -!- Malvolio [~Malvolio@idlerpg/player/Malvolio] has joined #hplusroadmap
00:59 < fenn> "I wonder how we'll be making chips in 2050" -> with a stone hammer on a flint nodule
01:03 < fenn> DNA origami tiles could aid the self assembly process, and the tiles could carry simple inorganic circuit components to their destination, where they get fused together eventually
01:03 < fenn> protein design is getting there too, lots of advances recently
01:04 < fenn> one could imagine a "mechanical" protein based computer
01:05 < fenn> also protein could make optical metamaterials for an optical computer
01:09 < maaku> drexler and merkle's calculations for mechanical atomic computers is pretty cool, but I don't see us ever moving off electrical substrates
02:48 < muurkha> lsneff: ultimately the appeal of reversible computing is that it repeals the Landauer limit
02:51 < muurkha> if you can do your computing 99% reversible, then you can run 50 steps forward, store the result, and then 50 steps back, and you've done 50 steps of computation for the energy cost of 2 irreversible steps
02:52 < muurkha> this is more important than it sounds because most of the limits on computation are ultimately energy limits
03:10 -!- TMM_ [hp@amanda.tmm.cx] has quit [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.]
03:10 -!- TMM_ [hp@amanda.tmm.cx] has joined #hplusroadmap
04:02 -!- sivoais [~zaki@199.19.225.239] has quit [Ping timeout: 260 seconds]
04:03 -!- sivoais [~zaki@199.19.225.239] has joined #hplusroadmap
04:09 -!- berndj-blackout [~berndj@197.189.254.139] has joined #hplusroadmap
04:09 -!- berndj [~berndj@197.189.254.139] has quit [Ping timeout: 256 seconds]
04:14 -!- berndj-blackout is now known as berndj
04:55 -!- yashgaroth [~ffffffff@2601:5c4:c780:6aa0::4249] has joined #hplusroadmap
04:59 < kanzure> instead of a protein mechanical computer you could also make a photomask with proteins
05:20 < nsh> complex molecules might be able to compute in far more interesting ways than by simulating relays
05:21 < nsh> computers aren't clever. they're just dumb at a very high density
05:21 < kanzure> https://twitter.com/zetalyrae/status/1620032247989690368
05:21 < kanzure> (orion's arm stuff)
06:26 -!- saturn2 [~visitant@user/clone-of-saturn/x-1551297] has quit [Ping timeout: 252 seconds]
06:43 < L29Ah> re lithography: https://www.nikon.com/products/semi/lineup/
06:44 < L29Ah> also mind that "<45 nanometer" products is most often bullshitometers
06:57 < L29Ah> i wonder how expensive are electron beam lithography setups
06:58 < lsneff> muurkha: I understand the benefits of reversible computing
07:07 < lsneff> nmz787: DSA is actually possible now? I never understood how people were planning on making structures that weren’t just a repeating pattern
07:07 < lsneff> I guess DSA could make the resist for the gates, and then metal layers could be patterned in a different manner
07:09 < lsneff> But chips don’t just have the same gate design repeated ad infinitum
08:01 < maaku>  
08:39 < muurkha> lsneff: What was it Mike said that discouraged you, then?
09:59 -!- cthlolo [~lorogue@77.33.23.154.dhcp.fibianet.dk] has joined #hplusroadmap
10:03 < hprmbridge> nmz787> lsneff yes DSA is a thing that's getting ready for high-volume mfg
10:04 < hprmbridge> nmz787> it isn't using DNA or proteins, rather simpler polymers often based on PMMA with a tuned length and tuned polarity at the tips/ends
10:04 < hprmbridge> nmz787> mostly the effort is being investigated for pitch division stuff
10:05 < hprmbridge> nmz787> https://en.wikipedia.org/wiki/Multiple_patterning#Directed_self-assembly_(DSA)
10:06 < hprmbridge> nmz787> basically expose resist with larger wavelength, then seed with DSA polymers which orient themselves around the exsiting resist islands/lines... then remove original resist, then only DSA 'resist
10:06 < hprmbridge> nmz787> remains
10:06 < hprmbridge> nmz787> which apparently increases the frequency of lines compared to the initial exposed pattern
10:22 < lsneff> Ah, that's disappointingly mundane
11:22 -!- TMM_ [hp@amanda.tmm.cx] has quit [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.]
11:22 -!- TMM_ [hp@amanda.tmm.cx] has joined #hplusroadmap
11:44 -!- cthlolo [~lorogue@77.33.23.154.dhcp.fibianet.dk] has quit [Read error: Connection reset by peer]
13:49 -!- Llamamoe [~Llamamoe@46.204.77.68.nat.umts.dynamic.t-mobile.pl] has joined #hplusroadmap
13:59 -!- Malvolio [~Malvolio@idlerpg/player/Malvolio] has quit [Ping timeout: 260 seconds]
14:03 < hprmbridge> nmz787> Not really, it could save your company $340mil on an EUV machine with high NA optics
14:04 -!- codaraxis [~codaraxis@user/codaraxis] has joined #hplusroadmap
14:42 < kanzure> diffusion for 3d objects and thingiverse https://twitter.com/KordingLab/status/1619885582905401344 (kinda)
14:59 -!- Llamamoe [~Llamamoe@46.204.77.68.nat.umts.dynamic.t-mobile.pl] has quit [Quit: Leaving.]
15:17 < fenn> very impressive, but note that NeRF (neural radiance fields) are not suitable for printing, and maybe not even photogrammetry (step needed to convert turntable images into a 3d model)
15:17 < fenn> it's not really a 3d model but generated images of an imaginary 3d model
15:17 < kanzure> well guess what, that's what you're going to get
15:18 < kanzure> let me predict the future: nobody will listen to you and all of our future 3d objects are going to be awful
15:18 < fenn> um, i mean it's the wrong format, and there are no converters
15:18 < fenn> how do you make a physical 3d object from that?
15:19 < L29Ah> s/physic/mathematic/
15:19 < fenn> even a crappy one
15:19 < kanzure> wasn't one of them a point cloud?
15:19 < fenn> yes the point cloud one would work. i was talking about dreamfusion and meta's make-a-video3d
15:20 < fenn> point-e is hot garbage
15:20 < juri_> hmm. i'm staring at 3d scanning problems atm.
15:20 < kanzure> what was the mathematical technique for taking a point cloud and getting a 3d object from it? was it "surface boundary regression"?
15:20 < L29Ah> https://dreamfusion3d.github.io/index.html this has "load 3d model" buttons
15:20 < L29Ah> i wonder how are they stored
15:20 < juri_> i might be writing some software in that area soon. i'm sure l29ah is just shivvering with antici...
15:20 < juri_> :)
15:21 < L29Ah> > Our generated NeRF models can be exported to meshes using the marching cubes algorithm for easy integration into 3D renderers or modeling software.
15:22 < juri_> oh wow.
15:22 < fenn> ok that would work
15:23 < kanzure> "point cloud segmentation by surface growing algorithm" uhh ok. i thought there was an analytical solution tho
15:24 < L29Ah> juri_: funnily i printed a 3d scanner but got too lazy to assemble it/reprint poorly printed parts since the software options i found were kinda eww
15:24 < L29Ah> (in fact a manual turntable)
15:25 < L29Ah> i wonder if a 3d printer robot can be abused nicely for 3d scanning
15:25 < L29Ah> duktape a camera on it and wee
15:25 < kanzure> all you need to do is find the bezier surface anchored to several million points that just so happens to match the point cloud
15:28 < fenn> perhaps someone could make a slicer that works on NeRFs, since finding the object's perimeter at a certain Z value is pretty similar to rendering
15:29 < fenn> the demo webGL models look okay
15:29 < fenn> i can't figure out how to download them to take a closer look in meshlab or whatever
15:30 < L29Ah> https://github.com/dreamfusion3d/dreamfusion3d.github.io/tree/main/assets/meshes2
15:30 < fenn> the NeRFs contain a lot more information about surface texture and details
15:40 < juri_> i think first, i need to know what format people are storing their scans in. i haven't really decided on a method, yet, but am just building a frame for motion controlled sensors. one of the things i'd like to detect is diffraction of light, as what i'm scanning is transparent.
15:40 < juri_> i'll see what i come up with, as i approach the target.
15:45 < fenn> there's "3d scanning spray" an opaque naphthalene coating that evaporates over time so as to not ruin the object being scanned
15:46 < fenn> perhaps you could use far UV imaging
15:47 < fenn> afaik scanning objects with reflections and refractions is an unsolved problem
15:52 < juri_> neat. thank you.
16:13 < fenn> the mesh quality isn't terrible: http://fennetic.net/irc/make-video3d_sweaterfrog_mesh.png
16:14 < kanzure> an anuran!
16:29 -!- Jay_Dugger [~jwd@47-185-201-55.dlls.tx.frontiernet.net] has joined #hplusroadmap
16:35 < fenn> probably usable for printing https://fennetic.net/irc/make-video3d_crab_mesh_cleaned_up.png
16:38 -!- TMM_ [hp@amanda.tmm.cx] has quit [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.]
16:38 -!- TMM_ [hp@136.243.23.165] has joined #hplusroadmap
17:02 < fenn> ok i should probably stop https://fennetic.net/irc/make-video3d_crab_mesh_cleaned_up2.png
18:15 -!- yashgaroth [~ffffffff@2601:5c4:c780:6aa0::4249] has quit [Quit: Leaving]
19:18 -!- Malvolio [~Malvolio@idlerpg/player/Malvolio] has joined #hplusroadmap
19:20 -!- Malvolio [~Malvolio@idlerpg/player/Malvolio] has quit [Read error: Connection reset by peer]
19:21 -!- Malvolio [~Malvolio@idlerpg/player/Malvolio] has joined #hplusroadmap
19:33 -!- codaraxis [~codaraxis@user/codaraxis] has quit [Ping timeout: 248 seconds]
22:11 -!- codaraxis [~codaraxis@user/codaraxis] has joined #hplusroadmap
23:10 -!- Llamamoe [~Llamamoe@46.204.77.68.nat.umts.dynamic.t-mobile.pl] has joined #hplusroadmap
23:14 -!- codaraxis__ [~codaraxis@user/codaraxis] has joined #hplusroadmap
23:17 -!- codaraxis [~codaraxis@user/codaraxis] has quit [Ping timeout: 255 seconds]
23:54 -!- codaraxis__ [~codaraxis@user/codaraxis] has quit [Ping timeout: 252 seconds]
--- Log closed Tue Jan 31 00:00:30 2023